AudioQA is a pipeline to answer questions based on a known context (user-fed) by speech or text. Users can speak their question and the system will return the answer as text. Future work will include TTS support as well.
| Argument | Info | Notes |
|---|---|---|
| --help, -h | Help for passing params | |
| --stt | Chooses stt model. Default silero | Supports silero, facebooks2t |
| --qa | Chooses model for question answering | Supports MobileBERT |
Models: Torch Hub and Transformers/HF
Package Development: https://packaging.python.org/en/latest/tutorials/packaging-projects/